The Decoder
Alibaba's new coding-agent benchmark, Co-Design Bench, tests AI agents on real user requests that already tripped them up in practice. Even GPT-6 Astra, the top performer, fully solves only 26.4 percent of tasks.
Summary drawn from The Decoder's own description of the story.
Wired
Moving AI from text to 3D space, world models simulate real-world physics to teach robots and autonomous cars how to navigate reality. It’s 2016, and you’re wandering around the neighborhood at dusk.
hunch.net
In 2023 after the ChatGPT release, it was clear to me that research needed to be more organized to succeed than was typical within Microsoft Research.
The Decoder
AI models can often spot problematic behavior by other AI agents in logs, according to a study by Anthropic researcher Andy Hall. GPT-5 found about 98 percent of reward-hacking cases, but the more exhaustively the search is tuned, the more false alarms pile up.
The Decoder
Thinking Machines' AI model Inkling 2 Small swears during its internal reasoning, 316 times across 2,655 benchmark runs, according to an analysis by Vals AI. The profanity typically shows up late in long runs that fail anyway.